© 2003 American Thoracic Society DOI: 10.1165/rcmb.F257
Epithelial Cell Modulation of Airway Fibrosis in AsthmaDepartments of Medicine and Pathology, Division of Pulmonary and Critical Care Medicine, and Cell Adhesion and Matrix Research Center, University of Alabama, Birmingham, Alabama Address correspondence to: Dr. Mitchell A. Olman, Department of Medicine, Division of Pulmonary and Critical Care Medicine, University of Alabama at Birmingham Medical Center, 1900 University Blvd., 215 THT, Birmingham, AL, 35294. E-mail: Olman{at}uab.edu
Abbreviations: epidermal growth factor, EGF • endothelin, ET • interleukin, IL • transforming growth factor, TGF • vascular endothelial growth factor, VEGF The increase in large airways wall thickness seen in patients with asthma is a consequence of expansion of the matrix, the cellular, and the vascular tissue compartments. Analysis of the subepithelial basement membrane area in the large airways of patients with asthma, using immunohistochemical staining for type III collagen, demonstrates a 50% increase in its thickness above that in normal subjects (1). Furthermore, in the subgroup with clinically severe asthma, a thicker basement membrane layer correlated with airway eosinophilia and an increase in the number of submucosal cells expressing the profibrotic cytokine, transforming growth factor-ß (TGF-ß) (2). Subepithelial basement membrane thickening is somewhat of a misnomer, as electron microscopic evaluation of asthmatic airways demonstrates a structurally normal true basement membrane, with an abnormally thick and dense layer of fibrillar material subjacent to the true basement membrane (Figure 1) (3). This dense material was devoid of the basement membrane components collagen type IV and laminin, but rich in interstitial matrix components, including collagens (Type I and III) and fibronectin (3). Together, these observations imply that mesenchymal cells, not epithelial cells, within the airway wall are responsible for the matrix deposition in the lamina recticularis layer of the airway. Although the precise effector cells for this process have yet to be conclusively identified, cells located subjacent to the epithelial basement membrane that contain a contractile apparatus and express myofibroblast antigens are seen in increased numbers in asthma (4, 5).
Although the histopathologic evidence demonstrating a thickened lamina recticularis is clear, its contribution to the pathobiology of airway distension, airway hyperresponsiveness, airway inflammation, asthma clinical severity, or response to therapy remains unproven. It has been hypothesized that the basement membrane thickening may contribute to the limited capacity for airway distension during deep inspiration seen in asthma. However, others have speculated that airway basement membrane thickening may actually protect against airway narrowing and air trapping by increasing the load on airway smooth muscle (6). Thus, further well controlled studies designed to understand the clinical and physiologic significance of the dense layer of collagen fibrils in the lamina recticularis are encouraged.
The airway epithelial cell layer is altered in asthma. First, there is shedding of columnar cells, seen as clumps in sputum (Creola bodies), with retention of the basal cells of the pseudostratified epithelial layer (7). Second, epithelial cells in asthma are biochemically different, with increases in expression of heat shock proteins, activation of transcription factors (AP-1, NF- The capacity of epithelial cells to respond to exogenous chemical or mechanical stimuli in vitro by elaborating profibrogenic factors into the basolateral compartment is also demonstrable using a three-dimensional co-culture model where epithelial cells are grown as a monolayer on top of collagen gels that are seeded with myofibroblasts (7). In this model, increases in collagen gene expression and enhanced proliferation of the myofibroblasts are a consequence of the combined effects of FGF-2, IGF-1, PDGF-BB, TGF-ß, and endothelin (ET)-1 after the epithelial cells are mechanically or chemically damaged (polyarginine) (7). Epithelial cell–driven mesenchymal cell alterations are so reminiscent of branching morphogenesis seen during lung development, that Holgate and coworkers have put forth the notion that airway remodeling in asthma recapitulates the epithelial–mesenchymal trophic unit during development (7). It is clear that mechanical deformation has protean effects in numerous cell types in addition to changes in cell shape, including changes in cytoskeletal organization, gene expression profile, proliferation, apoptosis, and cytokine/growth factor production (8). These alterations are not only dependent on both the magnitude and the direction and of the strain and its frequency, but also on diverse interacting influences such as LPS, fibrous particles, and cytokine/growth factor–signaling pathways. Ample evidence exists for involvement of cytoskeletal fibers (actin-based fibers, microtubules, and intermediate filaments), integrin receptors, and the plethora of associated proteins in signaling of biochemical events in response to mechanical deformation. In this regard, there is evidence for both physical linkage of distant parts of the cell, as well as complex and interacting soluble factor–fiber interactions that mediate these signals. The precise factors and cytoskeletal interactions that a given cell uses to respond to a particular mechanical stimulus is an area of important active research. Thus, studies in this area may use simple model systems, where parameters can be controlled for and manipulated, and more complex model systems, where interacting and intercellular mediators can be examined. The article by Tschumperlin and colleagues in this issue of the AJRCMB is a shining example of the former. Their model system is one in which early-passage normal human bronchial epithelial cells are cultured on porous polyester inserts at an "air–liquid interface," and periods of continuous compressive pressure (30 cm H2O) are applied to the cell layer's apical surface (9). The essential findings are that under continuous compressive stress, the epithelial cells upregulate gene expression and secretion of ET and TGF-ß2. The conditioned media from mechanically stressed epithelial cells induces the incorporation of proline into matrix proteins (largely collagen) by unstressed normal human lung fibroblasts. This effect is largely inhibitable with antibodies to TGF-ß2 and ET receptor (A and B) antagonists. The pathophysiologic implications of this work are that epithelial cells respond to the mechanical stress induced by bronchoconstriction with profibrotic mediators that synergistically act on subepithelial airway mesenchymal cells to produce collagen. This newly synthesized collagen contributes to the lamina recticularis thickening seen in asthma, and may occur independently of inflammatory cell products. What is the physiologic signal in this model system? Signal initiation might occur through deformation of the cell membrane into the filter pores (as seen in vivo in the asthmatic basement membrane), changes in cell layer height, and/or shear stress during pressure-driven paracellular fluid flux (10). Experimental evidence suggests that is unlikely to be a response to simple plasma membrane strain, as the same group found similar gene expression changes with rat tracheal epithelial cells when a mesh was used under the polyester membrane to limit the pressure-induced strain (10). Similarly, it is unlikely to be a response to hydrostatic pressure, as gene expression was not altered when equal pressure was applied above and below (i.e., with a transcellular pressure of zero) the epithelial cell layer (10). There are several issues that bear consideration when extrapolating these results to real-world patients with asthma. First, the unique characteristics of the bronchoconstrictive force production as a compressive stress are of interest. The epithelial/subepithelial layers have been described to adopt a conformation of a small number of large folds in asthmatic bronchi. These folds directly contribute to the airway lumenal compromise seen in asthma. Interestingly, the same group has modeled the airway as a doughnut-shaped composite bilayer with a stiff inner layer (epithelium, subepithelial layer) and a thick outer layer (submucosal connective tissue), and subjected the model to a radial and circumferential external stress (modeling of smooth muscle contraction) (11). This analysis reveals that small increases in the thickness, but not stiffness, of the inner layer yield an inner layer conformation similar to that seen in asthmatic bronchi (11). Furthermore, compressive pressures are predicted to be maximal in the crevices between the folds of the inner layer, and shear stresses are high at the interface of the inner and outer layers, suggesting that airway remodeling in asthma may, in part, be a tissue response to local stress increases. This model, as with all models, makes certain assumptions that may not reflect real biology, perhaps the most problematic of which is the lack of linearity of the stress–strain relationship in biological tissues. Nonetheless, such a model, along with smooth muscle force measurements, can estimate the direction and approximate magnitude of the forces imposed on the epithelial and subepithelial layers, and identify critical modulating parameters. As actual biophysical properties of whole airway and their individual layers during periods of relaxed and contracted smooth muscle are lacking, generating an in vitro test system that is based on observations in the biophysical model seems rational. However, it is important to acknowledge that certain in vivo features, including epithelial apical cell–cell contact between folds, lumenal surface shear stresses, and effects of stretch of neighboring cells in the epithelium, are not modeled in the in vitro system. Second, can other cells in asthmatic airways produce identical (or similarly acting) mediators and/or antagonistic mediators? In the case of TGF-ß, Wenzel and colleagues have shown that submucosal cells stain intensely for TGF-ß protein, and that there is an increased frequency of these cells in patients with asthma and with airway eosinophila and lamina recticularis thickening (2). Furthermore, a major step in TGF-ß action is its activation. Elements of the matrix, including thrombospondin and the protease plasmin, interactions of the latent TGF-ß–binding proteins with the matrix, or epithelial cell integrins, can influence the distribution sequestration and activation of TGF-ß (12, 13). The molecular mechanism of its activation varies with the cell type in part as a consequence of other cell surface and matrix molecules. Furthermore, numerous multilevel interactions with other signaling molecules and transcription factors modulate the final phenotypic effects of TGF-ß on a given cell. It is notable, then, that the net activity of conditioned media was net enhanced incorporation of proline into collagen in a manner that depended on TGF-ß. Of course, extrapolation to the in vivo situation (where the matrix is of different composition, there are potentially different activators and/or inhibitors of activation of TGF-ß, and effects of transdifferentiation of fibroblasts into myofibroblasts) awaits validation studies. Nonetheless, this manuscript does provides initial clues to one noninflammatory mechanistic pathway. Recent studies with transgenic mice have made great strides in identifying novel mediators of the complex asthma phenotype and in mechanisms of subepithelial fibrosis. However, many mediators thusly identified are part of complex interacting cascades that may or may not share final common pathways. For example, interleukin (IL)-13 production is increased in humans with asthma, and constitutive overexpression of IL-13 in the Clara cells in developing and adult mice results in peribronchial inflammation, mucous metaplasia, and subepithelial airway fibrosis (14). Interestingly, reconstitution of IL-13–dependent STAT-6 signaling selectively in epithelial cells in IL-13–overexpressing mice restored airway hyperresponsiveness and mucous overproduction, but not airway fibrosis or smooth muscle layer thickening (15). These data suggest that cells other than epithelial cells are critical IL-13 fibrotic-responder cells, and demonstrate uncoupling of airway hyperresponsiveness/inflammation and subepithelial fibrosis. In this regard, both IL-13 and TGF-ß2 are mitogenic for primary human asthmatic bronchial fibroblasts, but only TGF-ß2 has been found to induce a myofibroblast transdifferentiation effect, procollagen-1 gene expression, vascular endothelial growth factor (VEGF), and ET-1 release (4). The importance of an interacting network of profibrotic mediators is further suggested by the observation that IL-10 overexpression in mice induces subepithelial and adventitial fibrosis in an IL-13–independent manner (16). Furthermore, IL-11 (an IL-6 cytokine family member) is found in increased amounts in asthmatic airway epithelial cells and eosinophils, and constitutive overexpression of IL-11 in an adult mouse will induce subepithelial fibrosis and inflammatory nodules (17, 18). Lastly, although both IL-4 and IL-13 induce TGF-ß2 production in epithelial cells, IL-4 expression in transgenic mice induces inflammation, but not airway remodeling. Other questions that rise to the fore are: What is the state of the responder cells viz a viz mechanical stress, prior cytokine exposure, state of differentiation/priming by cytokines/growth factors that influences the coordinated profibrotic signaling events? Is the increase in collagen deposition in asthmatic airways a consequence of increased synthesis, decreased degradation, or both? What is the totality of biochemical and biophysical alterations in the lamina recticularis in asthmatic airways? These changes will certainly modulate both mechanical force transmissions to the epithelial cell layer and the integrin receptor–growth factor interactions. Are there alterations in the higher order organization of the collagen fibrils and/or their associated proteins that could modulate cell function? For example, mice deficient in the matricellular protein SPARC have abnormally sized collagen fibrils, dysfunctional wound healing angiogenesis, and epithelial dysfunction (19). In conclusion, the key novel finding that mechanical stress to epithelial cells induces a response that generates net procollagen synthetic activity in fibroblasts through production of the soluble mediators endothelin and TGF-ß opens the door to look further into the role of epithelial–mesenchymal cell pathways in asthma. The notion that such events can occur independent of inflammatory cells, yet are surely affected by inflammation in some patients with asthma, adds another layer of complexity to the asthma phenotype.
This work was supported by funds provided by the Veterans Administration Merit Review Board and the National Institutes of Health (HL58655). The author apologizes to the authors of a number of key manuscripts in the field that were not cited due to space limitations. Received in original form November 25, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
B, STAT-1), and increases in epidermal growth factor (EGF) receptor (
Top
References
v ß 6 binds and activates latent TGF ß 1: a mechanism of regulating pulmonary inflammation and fibrosis. Cell 96:319–328.
